The Extractive Logics of AI, Environmental Impacts and Other Challenges.

 

AI technologies are built and trained on vast amounts of online data. When Indigenous Peoples’languages, knowledge, and cultural materials are included in such datasets without transparency and their FPIC, it risks perpetuating patterns of exploitation and appropriation that Indigenous Peoples long have resisted. The data that AI models are trained on also frequently excludes or misrepresents Indigenous Peoples, their knowledge and voices. AI algorithms also tend to be biased by the worldviewof the developers. Such AI models are thus likely to reflect and may even exacerbate existing inequities. For example, with increasing use of biometric and facial recognition technologies, this can contribute to further misidentification and profiling of Indigenous Peoples. Moreover, AI systems depend on immense computational infrastructure with data centers that require significant amounts of electricity for their operations, water for cooling, raw materials for manufacturing electronics. This can significantly intensify climate and environmental pressures. When situated near Indigenous Peoples’ territories and lands, they can also exacerbate environmental degradation and resource scarcity, negatively affecting water availability and fragile ecosystems that Indigenous Peoples depend on for their survival and ways of life. Construction of data centers and manufacturing of the electronics needed also requires significant raw materials, critical minerals and rare elements. Mineral extraction frequently results in land dispossession, environmental degradation, loss of livelihoods, and threaten the health, cultural heritage and spiritual connections of Indigenous Peoples whose land are rich in extractive minerals. The additional demand for critical minerals and rare elements from such data centers risks exacerbating such pressures and impacts. For example, in northern Chile Indigenous Atacameño Peoples are resisting AI-powered mining operations that extract lithium and copper—key materials for AI hardware and electric vehicles. These operations use AI to optimize extraction and logistics, threaten sacred lands and water sources in the Atacama Desert and are being challenged for violating Indigenous rights and environmental protections.There are also risks tied to the electronic waste produced by such data centers, which often contain hazardous substances such as mercury and lead. Mercury exposure linked to extractive industries already presents a health crisis that negatively affects Indigenous Peoples, and in particular, Indigenous women. As such, e-waste from AI data centers could further expose Indigenous Peoples to toxic substances when discarded on or near their lands.